Measurements on a Tube Radio

Peter D. Hiscocks

Professor Emeritus, Electrical Engineering, Ryerson University

phiscock@ee.ryerson.ca

March 12, 2011

Contents1 Introduction 2

1.1 Isolation Transformer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.2 Caution: Lethal Voltages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31.3 Thermal Shock and the Adjustable Line Transformer . . . . . . . . . . . . . . . . . . . . . . . 31.4 Being Paranoid: First Time Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 The All American Five AM Radio 42.1 The Schematic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.2 Chassis Connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3 Input, Rectifier and First Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.4 Second Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.5 Audio Power Amplifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.6 IF Amplifier Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92.7 Local Oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92.8 IF Amplifier Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102.9 All American Five Schematic: Original . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

References 11

List of Figures1 Isolation Transformer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Line Voltage Autotransformer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Nipper II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Redrawn Schematic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Power Supply Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Audio Power Amplifier. Upper trace, Plate voltage, 20V/div. Lower trace grid voltage, 2 volts/div 77 Amplifier Frequency Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Output Transformer Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 IF Amplifier Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 910 Local Oscillator Waveforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 911 IF Amplifier Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1012 Original Schematic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

1

1 IntroductionIn this paper we show how a modern measuring instrument, the Syscomp CGR-101 oscilloscope-generator, canbe used to view the operation of vacuum tube based equipment. We give an example, showing measurementson an All American Five tube radio, the Nipper II, made by RCA in Montreal, Canada. Modern instruments arevery useful in diagnosing faults in vacuum tube based equipment, but there are some precautions that need to beobserved for the safety of the human being, and the preservation of the equipment.

1.1 Isolation TransformerThe All-American-Five has no transformer isolation from the AC power line. The electrical common connectionin the radio, which connects to the metal chassis, is connected to one side of the AC line cord. The power cordon the AA5 is not polarized it can be plugged into an AC outlet in two different ways. For one configuration,the radio common (and chassis) are connected to the AC line neutral, which is within a few volts of earth ground.For the other configuration, the circuit common and chassis are at AC line potential, 117VAC.

Figure 1: Isolation Transformer

Now consider what happens when the AA5common is at 117VAC and you connect an in-strument ground (such as a CircuitGear oscil-loscope probe) to the radio circuit common. Ineffect, there is a dead short across the AC line.This will put a momentary surge of tens of am-peres through the circuit, vapourizing sectionsof circuitry and components. (Ultimately, thefuse breaker in the distribution panel may open,but that will be far too late to prevent damage.)

Some people suggest floating the oscillo-scope off ground. For example, if the host lap-top is battery powered, then one can theoret-ically connect the oscilloscope ground to theradio ground. There will no longer be a deadshort across the AC line, but any exposed metalparts of the oscilloscope and laptop can possi-bly be at a potential of 117VAC. A human op-erator is at great risk of serious shock or elec-trocution.

The safe way to work on such a radio is to power it from an isolation transformer. This is a 1:1 transformerthat delivers 117VAC to the radio from a secondary that is totally isolated from ground. Consequently, eitherterminal of the secondary can be connected to an earth ground without risk.

One can buy a ready-made isolation transformer, for example, one of the the 171 series from HammondManufacturing 1. The 100VA model is available from Digikey for $100. The 171 series is ready to go, with aprimary circuit breaker, primary plug and secondary outlet wired in.

You can also build your own isolation transformer, as shown in figure 1. I used the Signal Transformer A41-80-230, an 80VA unit available from Digikey for $34. The AA5 radio requires 0.3 amps at 117V, or 35VA, so thistransformer has an adequate power rating. The A41-80-230 transformer has dual primaries and dual secondaries.For 117VAC operation, the primaries are wired in parallel. Check the data sheet carefully: the dotted terminalsmust be connected together.

1http://www.hammondmfg.com/pdf/PB0231.pdf

2

The circuit is otherwise very straightforward: the primary is connected to the AC line, in series with a fuseand on-off switch. The secondary is connected to an AC outlet, which is now the source of isolated AC voltage.

A cautionary note: an autotransformer consists of one transformer winding tapped at various points to stepup or step down the input voltage. A common example of autotransformer is the Variac adjustable transformer,which can be used to adjust line voltage. An autotransformer does not provide isolation from the AC line. If youwant adjustable line voltage and isolation, you need both an autotransformer and an isolation transformer.

1.2 Caution: Lethal VoltagesModern solid-state equipment uses supply voltages typically in the range of 3 to 15 volts. Tube equipment supplyvoltages start at 100 volts and go up from there. The supply voltages in a tube power amp may be something like+350 volts. In an RF linear amplifier or microwave oven, thousands of volts are not unusual. The casual approachto safety that works for a 3.3 volt op-amp could result in deadly harm from a vacuum tube based unit.

The radio used as an example in this note has a relatively modest supply voltage - under 170 volts. However,this is still high enough to cause a nasty shock or even, under the right circumstances, cardiac arrest. The isolationtransformer minimizes the chance for a supply-to-earth-ground shock, but it can still happen if one hand is on thechassis common and the other touches a high voltage.

Work slowly, deliberately and carefully.

Shut off the power when making changes to the circuitry or measuring configuration.

When using the oscilloscope to measure a signal, first use a DC multimeter (such as the Syscomp DVM-101) to measure the DC voltage at that point. If you need to observe a small AC voltage riding on top ofa large DC value, block the DC value with a capacitor. Make sure the capacitor has the necessary voltagerating.

If the scope probe has a switchable 1/ 10 setting, start the measurement in the 10 position.

When injecting a signal from the generator, bear in mind that the low internal resistance of the generatorwill short out any DC or AC voltage at that point in the circuit. Use a blocking capacitor to prevent the DCfrom flowing through the generator.

1.3 Thermal Shock and the Adjustable Line Transformer

Figure 2: Line Voltage Autotransformer

Vacuum tubes are heated by a filament. Theresistance of the filament is low when the fil-ament is cold. As a result, there is a large in-rush of current through the filaments when theradio is first switched on, especially if the mo-ment of switch-on coincides to a peak in the ACline voltage. (The same effect is at work withincandescent lamps, which tend to burn out atthe time they are switched on.) There may besome mechanism in the radio to minimize thisinrush current. If not, or if you wish to be cau-tious, you should apply line voltage from anadjustable transformer (figure 2). Turn on thepower switches and increase the supply voltageup to 117 volts.

3

1.4 Being Paranoid: First Time OperationThe first time you operate the radio from the isolated supply, its a good idea to measure the isolation between theradio chassis and the ground point of the measuring instrument. With the radio plugged into its power source (theisolation transformer) and power off, use an ohmmeter to measure the resistance between the radio chassis andthe measuring instrument ground (such as the barrel of a BNC connector). The reading should show open circuit.

Now enable power to the radio. With a voltmeter on the AC voltage range, measure the potential between aground point on the oscilloscope and the chassis of the radio. The voltmeter should read zero volts.

If the system passes both these tests, its safe to work on.

2 The All American Five AM RadioThe All American Five (AA5) is a 5 tube AM band radio that was produced in great numbers for the consumermarket in North America. The background, history and some technical information on the AA5 is availablehere: [1], [2], [3]. Description and technical information is in RCA Receiving Tube Manuals, [4] [5].

This particular radio has a label on the bottom of the case which states the requirement for a license, accordingto the Radio Telegraph Act (Canada) of 1938. Mandatory licensing of private receiving stations was abolishedin 1952, [6] so this set was sold some time between those two dates. Its still operating properly in 2011, some 60years later.

(a) Case (b) Chassis Front

(c) Chassis Top (d) Chassis Wiring

Figure 3: Nipper II

4

2.1 The SchematicIts often helpful when trying to understand some new circuit to redraw the diagram from scratch. Ideally, themost positive power supply voltage should be at the top of the page, and the signal should flow from left to right.Compare the original schematic of figure 12 on page 11 with the redrawn schematic of figure 4 on page 5. Itsmuch simpler to trace the power supply wiring and some of the signal connections on the redrawn version.

Throughout our exploration of the radio operation, we will refer to the redrawn schematic.

30u150V

34

34

34

34

34

1

2

34

5

6

7

Tube BaseBottom View

T4

150p

3M3

150p

47k

56p

33k

217pF

7 115pF

217pF

217pF

Loop Antenna

4M7

500k

20n400V

330p

1k2, 1W

35W4

50C5

12BA6

12BE6

12AV6

117VAC

50uF150V

100n400V

AVC

12BE6 12BA6DetectorPentagrid Converter IF Amplifier

AVCAudio Amplifier

12AV6

Audio Amplifier50C5

Rectificer35W4

100R

2nF, 400V

2nF, 400V

470k

220kR1C4

5

C16

R11

C8

7

7

62

1

150RR10

R9470k

C12

S1

2

1

6

50n, 50VC2

Figure 4: Redrawn Schematic

2.2 Chassis ConnectionThe metal chassis is connected to the power supply common (which is one side of the AC line!) via 220k resistorR1 in parallel with 100nF capacitor C4. It would be convenient to use the chassis as the ground point for ourmeasurements. Then this resistor and capacitor will appear in series with our measurement probles. Will thatwork?

The input resistance of the Syscomp DVM-101 digital voltmeter is 10M. Consequently, the 0.22M resis-tance will introduce a 2.2% loading error, which is negligible in this application.

The Syscomp CGR-101 CircuitGear oscilloscope has a maximum input range of 25 volts, well below whatis required to measure the AA5 supply voltages. We will need to use a 10 oscilloscope probe, which extendsthe range to 250 volts. Again, the 0.22M resistance has very little effect in this application.

What is the effect of the bypass capacitor C4? Consulting a reactance slide rule, the capacitance is equal tothe resistance at 7Hz. Above that frequency, its impedance is less than the resistance. At the 60Hz power supply

5

(a) Input & Output (b) Output & 2nd Filter

Figure 5: Power Supply Measurements

frequency, the impedance will be roughly 25k. At the RF frequencies of the radio (540 to 1600 kHz), thiscapacitance will appear essentially as a short circuit. In either case, the capacitor impedance is much less than theinstrument impedance and it will not affect the voltmeter or oscilloscope.

So: The measuring instrument ground (voltmeter or oscilloscope) can be connected to the chassis.

2.3 Input, Rectifier and First FilterReferring to the schematic of figure 4, the line voltage is connected to the 35W4 half wave rectifier. The rectifierconducts on positive half cycles of the line voltage. These pulses are filtered first in 50F capacitor C16. Thewaveforms of the input voltage and capacitor voltage are shown in figure 5(a)2.

There are several interesting features in these waveforms:

The supply voltage has a peak value of around 186 volts.

The voltage drop across the rectifier diode, when conducting, is about 3 minor divisions on the screen, or 30volts. That compares to about 0.7 volts for a modern silicon rectifier diode. (This is a reason to be careful inreplacing a tube type diode with a semiconductor diode. The output voltage will be larger and may exceedthe voltage rating of other components.)

The input sine wave is slightly distorted during the conduction interval of the diode. The current flow duringthat interval creates a voltage drop in the isolation transformer.

2Figure 5(a) shows the complete GUI (graphical user interface) for the Syscomp CircuitGear oscilloscope used in these measurements.Figure 5(b) and subsequent figures just show the waveform display area.

6

The peak to peak ripple across the filter capacitor is about 30V. Knowing the time interval of the dischargeand the change in voltage, one could calculate the current drain in that section of the power supply.Alternatively, one could measure the current at that point using an ammeter and then calculate the actualfilter capacitance.

The filtered power supply voltage is used to power the 50C5 audio power amplifier stage. One might reason-ably expect the 30 volt ripple on the supply voltage to show up as a nasty hum in the loudspeaker. However, the50C5 is a pentode (ie, 5 element) tube, and so the output is relatively high incremental output impedance. In otherwords, the pentode behaves as a voltage-controlled current source, and the value of the plate voltage has littleeffect on the plate current. Consequently, the ripple is largely ignored.

2.4 Second Filter

Figure 6: Audio Power Amplifier. Upper trace, Plate voltage,20V/div. Lower trace grid voltage, 2 volts/div

The rectified and filtered power supply voltageis then further smoothed in RC lowpass filterR11 and C8. The input and output voltages arecompared in figure figure 5(b).

There is a 50 volt drop between the av-erage input voltage to this filter and theoutput.

This supply voltage, which is used for theremaining stages of the radio, is about 91volts.

The R11, C8 filter is very effective inremoving ripple from this much lowersupply voltage. The peak-peak ripple isabout one minor division, or 4 volts.

2.5 Audio Power AmplifierThe input to the audio power amplifier stage is the grid terminal at pin 2. The output is on the plate terminal atpin 7. Figure 6 shows the signals on the two terminals.

The top trace is the plate voltage which shows some of the low frequency undulation due to the ripple on thepower supply voltage, as mentioned earlier. The smaller undulations are the audio signal.

The lower trace is the grid voltage. Notice that the grid voltage signal and plate voltage signal are out of phase:as one increases, the other decreases.

Comparing the amplitudes of the two audio signals and allowing for the scale factors, the amplifier voltagegain is approximately 5 volts/volt.

Now look at the positioning of the zero voltage points for both traces. Zero voltage for the plate signal is atthe bottom of the screen (the small B to the right of the display area), indicating that the signal is sitting on a largepositive DC voltage.

Zero voltage for the grid signal is above the trace, indicating that the signal is sitting on a negative voltage.This negative voltage is caused by electrons that hit the grid while flowing from cathode to plate in the tube. Theelectrons flow out of the grid negative voltage and establish a negative voltage across the grid resistor R9.

7

Figure 7: Amplifier Frequency Response

(Alternatively, think of current flowing into the grid,which makes the grid terminal negative with respect toground.) This current is known as grid leak current3.

The cathode terminal at pin 1 is biased to about +10volts, which is supposed to make the grid-cathode volt-age negative by the same amount.

Amplifier Frequency Response

The audio amplifier frequency response was measuredusing the network analyser function of the CGR-101.The result is shown in figure 7. This is the relative ampli-tude of signals between the input to the volume controland the output across the loudspeaker.

The input signal was applied from the CGR-101 gen-erator output to the top of the volume control. The oscil-loscope channel A, which is the reference input for thenetwork analyser function, was connected to the samepoint. The oscilloscope channel B was connected acrossthe loudspeaker terminals. The voltage at the top of the volume control is the negative half-wave rectified signalfrom the previous intermediate-frequency amplifier. As a result, the top terminal of the valume control is sittingat a negative potential that depends on the magnitude of the received radio signal. If we connect the generatoroutput of the CGR-101 to that point, the low internal resistance of the generator (150) will shunt that voltage toground. This is not particularly harmful, but just to be safe we put a 680nF DC blocking capacitor between theoutput of the generator and the volume control terminal.

Figure 8: Output Transformer Ratio

The frequency response shows a 5db rise, peakingaround 4KHz. This may or may not be reflected in theactual acoustic output, which depends on the response ofthe loudspeaker. In any case, no one would mistake thisfor a high-fidelity device.

Audio Output Transformer

Curious about the turns ratio of the audio output trans-former, I injected a signal into the volume control andmeasured the voltage at the primary (red trace) and sec-ondary (blue trace), as shown in figure 8.

The turns ratio is about N = 20 : 1. This is usuallyexpressed as the ratio of impedances, which is N2 =400. Then an 8 ohm speaker would appear as 3200 ohmsin the primary of the transformer.

3When first measured, the grid voltage was sitting about +9.5 volts. Looking at the schematic, almost no current should be flowing inthe grid resistor R9, and the voltage should be around zero. The fact that it was positive suggested that C12 is leaking current from the verylarge voltage at the plate of the preceeding 12AV6. Capacitors do tend to deteriorate with time, and this one is over 60 years old, so itsnot surprising that it would be come leaky. Replacing the capacitor stopped the leakage current and allowed the audio volume to increasesubstantially.

8

2.6 IF Amplifier Output

Figure 9: IF Amplifier Output

The IF (intermediate frequency) signal should consist ofan RF carrier at the IF centre frequency (455kHz), ampli-tude modulated by the audio signal. This isnt clearly ev-ident from an oscilloscope waveform trace, but it showsup nicely on a spectrum display (figure 9). The trace nearthe cursor is the IF centre frequency at 455 kHz. Noticethat there is some slightly observable width to this sig-nal, caused by the audio modulation. The low-frequencyspectrum is the demodulated audio. The other spectrumspikes are other radio stations.

As the radio is tuned, the radio stations move left orright on the spectrum display, giving a simple panadaptorreadout.

2.7 Local Oscillator

(a) Cathode (Pin 2) (b) Grid (Pin 1)

Figure 10: Local Oscillator Waveforms

In a superheterodyne radio such as this one, the RF tuning and the local oscillator track each other so that the localoscillator frequency is different from the signal frequency by a fixed amount, the IF frequency. The two signalsare mixed (multiplied) and the output frequency is then fed into the IF (intermediate frequency) amplifier. As aconsequence, the selective amplification of the signal can take place at a fixed frequency.

The tuning of the RF circuit and the local oscillator is by two air-variable capacitors, which can be seen infigure 3(c).

9

In AM band radios, the input frequency tunes over the AM broadcast band, 540 to 1600 kHz (or kilocycles, asit used to be called). The intermediate frequency is 455 kHz. The local oscillator (LO) uses high side injection, soit is always 455 kHz higher than the input frequency. Consequently, we should see the frequency of the LO varybetween 995kHz and 2055kHz.

The local oscillator is the transformer-variable capacitor network connected between the cathode (pin 2) andfirst grid (pin 1) of the 12BE6 Pentagrid Converter. The waveforms are shown in figure 10.

The cathode waveform shows a series of pulses. As the radio is tuned, their frequency varies from 987 to2100kHz. This measurement was taken with the probe on the x10 setting, so that it presents 10M in parallelwith approximately 2 picofarads of capacitance. The cathode of the tube is a relatively low impedance point, sothat helps to minimize the effect of attaching the scope probe. However, radio frequency circuits are exquisitelysensitive to loading, and the probe has enough of an effect to shift the frequency of operation slightly. Changingthe probe to its x1 setting causes a much larger error in the measured frequencies.

The grid waveform shows a sine wave at the local oscillator frequency. The tuned network in the localoscillator has removed the higher order harmonics so that the injected signal into the mixer is relatively clean.This is a higher impedance point than the cathode and the probe has a substantial effect on frequency.

2.8 IF Amplifier Response

0

50

100

150

200

250

300

350

425 430 435 440 445 450 455 460Frequency, kHz

Amplitude, mV

Figure 11: IF Amplifier Response

For best operation, the IF frequency should be cen-tred on 455kHz, have the proper bandwidth, and besymmetrical in shape. At the time this radio wasconstructed, the IF response was measured by whatthe British called a wobbulator: an RF signal gener-ator that could be swept over a range of frequencies.The output of the IF amplifier was sent to the ver-tical plates of an oscilloscope. At the same time,the signal sweeping the RF was applied to the hor-izontal plates of the scope. The result was an am-plitude vs frequency display of the IF response. Inorder to calibrate the display, there were frequencymarkers generated by crystal frequency standardsthat showed up as bumps on the display4.

That technique is not possible with the CGR-101, but there is an alternative. The CGR-101 os-cillator frequency can be set to a precision of 0.1Hz,so the RF input to the radio can be set very precisely.This can be used to sweep the IF frequency over arange, while measuring the output voltage from theIF amplifier. The exact IF frequency of a waveform can be measured quickly and accurately using the CGR-101oscilloscope Automeasure facility.

The CGR-101 generator output was connected to the antenna terminal of the radio. Channel A of the CGR-101 scope was connected to the IF amplifier output, at pin 6 of the 12AV6. The automatic level control loop wasdisabled by short-circuiting capacitor C35.

Referring to figure 5, there are some features of the CGR-101 generator that make this measurement straight-forward. The labels at the top and the bottom of the frequency slider are clickable. When mouse-clicked, an entry

4The author recalls visiting a radio factory in Toronto in the 70s, where a row of workers adjusted the responses according to a grease-pencil template on the oscilloscope screen.

5This appeared to have no effect on the amplitude of the signal, so it may be that the ALC is defective on this radio.

10

widget opens up and one can enter a new value. You can narrow down the range of the frequency slider to theRF frequency of the radio: 540kHz to 1600kHz. As well, the frequency readout display is clickable and one canenter a specific frequency.

For this measurement, set the frequency of the radio to some spot on the dial where there are no major radiostations. Slowly sweep the radio frequency over its range, watching the signal amplitude on the oscilloscope. Atsome point, you will see an increase in the signal.

Zero in the frequency limits to bracket the region where the IF amplitude ranges from a small value, tomaximum, back to small amplitude. This is the region of measurement.

Step the frequency through this region. At each frequency, use the Automeasure facility of the scope tomeasure the IF frequency. Use the amplitude cursors to measure the amplitude of the sine wave.

Figure 11 shows the result for this radio. The response peaks at 448kHz rather than 455kHz and is ratherunsymmetrical. It could probably be improved. However, before moving the centre of the IF, it would be wise tocheck on the relationship of the local oscillator and signal frequencies. It may be that the LO is offset by 448kHz,in which case the radio is tuned correctly. Re-aligning an IF amplifier is not for the faint hearted - adjustmentsrequire special tools, a non-magnetic slug adjustment tool for the IF transformers and a non-magnetic miniaturescrewdriver for the capacitors. Furthermore, all the adjustments interact. You need to be able to read out thefrequency of the local oscillator accurately, without affecting it.

2.9 All American Five Schematic: OriginalFigure 12 is a copy of the original schematic, from reference [4].

Figure 12: Original Schematic

11

References[1] All American Five

http://en.wikipedia.org/wiki/All_American_Five

[2] History of the AA5 (All American 5ive) AM tube radiohttp://home.netcom.com/~wa2ise/radios/aa5h.html

[3] The All American FiveMax Robinsonhttp://www.angelfire.com/electronic/funwithtubes/AA5-1.html

[4] RCA Receiving Tube ManualTechnical Series RC-22RCA Victor Company Limited, 1963Available through used book sellers, such as the Academic Book Exchange: http://www.abebooks.com

[5] RCA Receiving Tube ManualTechnical Series RC-25RCA Victor Company Limited, 1966Available through used book sellers, such as the Academic Book Exchange: http://www.abebooks.com

[6] Radio Telegraphy Act in CanadaRadio Act Consolidationhttp://my.tbaytel.net/va3rom/docs/Radio%20Act%20Consolidation.pdf

12